Attached is a patch against 2.4.5-ac6 which should fix issues with
on-board Via devices and interrupts, particularly on Via SMP machines,
or Via machines that requires 'noapic' or "PNP-OS:No" to be functional.

This has been tested by a user with a Via SMP board under both MPS 1.1
and MPS 1.4, and has been shown to solve the problems he was having with
interrupt delivery.  It has also been tested on my dinky, ancient K6-2
laptop to make sure there were no generic Via regressions.

Special thanks again to Adrian Cox, who pointed out the
somewhat-documented fact that PCI_INTERRUPT_LINE is not merely a
scratchpad register for the on-board Via devices.  Since we blindly
write the values from IO-APIC into PCI_INTERRUPT_LINE, whose irq values
can exceed the register limit of 14, this caused some chaos for Via
users.

This patch also moves the Via ACPI to drivers/pci/quirks.c, because Via
ACPI is not x86 only, and it updates some comments and adds a helpful
printk.

2.5/LATE 2.4 NOTE:  As Linus has pointed out, we really need
infrastructure to have different interrupt routers on a per-device
basis.  If you look through the PCI quirks in the arch and generic code,
you see code working around this lack of infrastructure, sometimes in
really ugly ways.

-- 
Jeff Garzik      | Echelon words of the day, from The Register:
Building 1024    | FRU Lebed HALO Spetznaz Al Amn al-Askari Glock 26 
MandrakeSoft     | Steak Knife Kill the President anarchy echelon
                 | nuclear assassinate Roswell Waco World Trade Center